Is it possible to build TensorRT engine for given compute capability (different from device it is running)?

mark.eremeev · February 5, 2021, 5:01am

I want to build engine for GPU with compute capability 6 on platform having GPU with compute capability 7.5. Is it possible? if yes how?

spolisetty · February 8, 2021, 5:43pm

Following links may help you.
Installation guide for TensorRT

Developer guide to use TensorRT

Thank you.

Topic		Replies	Views
min GPU requirements to run tensorRT? TensorRT	2	4741	October 12, 2021
How can we build a tensorRT model just once and run on different GPUs? TensorRT	3	406	May 5, 2020
Cross-compilation TensorRT serialized engines TensorRT	2	916	October 12, 2021
TRT engine across different driver version TensorRT	3	1443	December 1, 2020
Is there any way to run a tensorrt engine without tensorrt environment? In another word, what is the minimal trt runtime environment? TensorRT	1	378	March 3, 2023
.engine generated on device A can`t be deployed to device B TensorRT tensorrt	3	391	May 17, 2023
Build TensorRT on Cuda compute capability 7.5 and make it backward compatible with previous capabilities TensorRT tensorrt	4	1950	May 19, 2022
Tensor rt confused about version TensorRT	1	422	July 6, 2020
The engine plan file is generated on an incompatible device expecting compute 6.1 got compute 8.6,please rebuild TensorRT	3	4092	April 19, 2024
Trtexec generates different engines when using the same platform/machine with the same onnx model TensorRT	3	1145	March 29, 2022